https://github.com/shaundann/autosight
AutoSight is an AI-powered multi-agent data analysis pipeline built on Google Cloud. From ingesting raw CSVs to generating visualizations and natural language summaries — all results are displayed live in a Streamlit dashboard.
https://github.com/shaundann/autosight
ai-agents automated-data-analysis bigquery data-pipeline gcp google-cloud llm multi-agent-systems python streamlit vertex-ai
Last synced: 9 months ago
JSON representation
AutoSight is an AI-powered multi-agent data analysis pipeline built on Google Cloud. From ingesting raw CSVs to generating visualizations and natural language summaries — all results are displayed live in a Streamlit dashboard.
- Host: GitHub
- URL: https://github.com/shaundann/autosight
- Owner: shaundann
- Created: 2025-07-10T18:01:58.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-07-10T18:11:54.000Z (9 months ago)
- Last Synced: 2025-07-11T00:44:08.822Z (9 months ago)
- Topics: ai-agents, automated-data-analysis, bigquery, data-pipeline, gcp, google-cloud, llm, multi-agent-systems, python, streamlit, vertex-ai
- Language: Python
- Homepage:
- Size: 26.4 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# 🤖 AutoSight – Automated AI Data Analysis Pipeline on Google Cloud
**AutoSight** is a multi-agent AI system that automates the entire data analytics pipeline — from dataset ingestion to AI-powered insights — using Google Cloud Platform (GCP).
With just a CSV URL, AutoSight:
- 📥 Crawls and uploads the dataset to Google Cloud Storage
- 📊 Cleans and analyzes trends using pandas and seaborn
- 🧠 Generates summaries using Gemini or GPT-like models
- 🗃️ Loads data into BigQuery for structured querying
- 📈 Displays all results in a modern Streamlit dashboard
---
## 🚀 Features
- 🌐 Web-based CSV ingestion from any public URL
- 📈 Trend analysis + visual plots
- 🤖 Natural language summaries with LLMs
- ☁️ Cloud-native: Cloud Storage + BigQuery
- 🖥️ Streamlit-powered live dashboard
- 💡 Modular agent architecture for easy extensibility
---
## 🛠️ Tech Stack
| Layer | Technology |
|---------------|----------------------------------|
| 💻 Backend | Python 3.12 |
| ☁️ Cloud | Google Cloud Storage, BigQuery, Vertex AI |
| 🤖 AI | Gemini Pro (or OpenAI GPT) |
| 📊 Visualization | matplotlib, seaborn, Streamlit |
| 🧱 Architecture | Multi-Agent System |
---
## 📁 Folder Structure
```
AutoSight/
├── main.py # Orchestrates entire pipeline
├── dashboard.py # Streamlit frontend
├── requirements.txt # Python dependencies
├── README.md # This file
│
├── agents/
│ ├── data\_crawler\_agent/
│ │ └── agent.py # Downloads CSV to GCS
│ ├── analyzer\_agent/
│ │ └── agent.py # Analyzes data + generates summary
│ └── bigquery\_writer\_agent/
│ └── agent.py # Loads dataset into BigQuery
```
---
## ⚙️ Setup Guide
### 1. Clone the repository
```bash
git clone https://github.com/your-username/AutoSight.git
cd AutoSight
````
### 2. Create a virtual environment
```bash
python3 -m venv venv
source venv/bin/activate
```
### 3. Install dependencies
```bash
pip install -r requirements.txt
```
### 4. Configure GCP
1. Enable these APIs:
* Vertex AI
* Cloud Storage
* BigQuery
2. Create a service account with roles:
* `Storage Admin`
* `BigQuery Data Editor`
* `Vertex AI User`
3. Download its credentials file as:
```
autosight-agent-key.json
```
### 5. Run the pipeline
```bash
python main.py
```
The script will:
* Download and store the dataset in GCS
* Analyze and generate a plot + summary
* Load structured data into BigQuery
* Launch the interactive dashboard automatically
---
## 📊 Example Dataset
Default dataset:
> [Air Travel (1958–1960)](https://people.sc.fsu.edu/~jburkardt/data/csv/airtravel.csv)
---
## 🧠 Future Enhancements
* [ ] Upload custom CSVs via dashboard
* [ ] Deploy Streamlit app to GCP App Engine
* [ ] Add forecasting and clustering agents
* [ ] Multi-dataset support + scheduling
---
## 🙅♂️ .gitignore
Create a `.gitignore` with the following:
```
autosight-agent-key.json
output.png
summary.txt
__pycache__/
*.pyc
venv/
.env
```
---
## 💬 Contact
Built by **Shaun Danny**
📧 [shaundanny2007@gmail.com](mailto:shaundanny2007@gmail.com)
💼 [LinkedIn]((https://www.linkedin.com/in/shaundanny/))
---
## 🏁 Acknowledgements
Inspired by [Google Cloud Multi-Agent Hackathon](https://googlecloudmultiagents.devpost.com) and the [original InsightAgents](https://github.com/Soulfullmens/insightagents) project.
```
---
Would you like me to create and export this `README.md` file for you directly so you can drop it into your GitHub repo?
```